The investor viewpoint
Users of the Takahashi-Alexander model or a variation of it have tended to be more sophisticated LPs, according to Patrick Sherwood, principal at GroveStreet and a former investment professional at the Yale
Investment Office, where he worked alongside Takahashi. “Many others have tended to use a rule of thumb, where they assume that the capital will be called over the next five years,” he says. “Other teams will attempt to generate a bottom-up forecast based on input from GPs, but that is quite labour-intensive and relies on the investor being in close contact with the fund manager – that’s not always possible and it’s also subject to human bias.”
Takahashi-Alexander, therefore, has been “the best approach available”, adds Sherwood. “The challenge is not to treat it as 100% correct. It’s more of a general guide to cash flows, yet there can be a temptation to assume that the number it produces is a prediction of what is going to happen.” This is particularly true, he says, when the results are shared with audiences across or outside an organisation, as they may not know the model’s limits and may become attached to a particular forecast number. As a result, Sherwood agrees with the paper’s assessment that Takahashi-Alexander’s main limitations are that it requires assumptions as inputs and that it produces a single outcome. “It’s not clear that everyone always understands the inputs used for the model,” he says. “When Yale first adopted this, it was looking back 20 years to the 1980s – that’s very different to today. The market has evolved and changed so much over the past 10 to 20 years – it has grown and there are many different types of funds. You can’t use this blindly.”
So what about the new model put forward by the researchers at Bella? “It’s a tremendous improvement to be able to ground the future in a range of possibilities based on real data – it’s a true advance,” he says. One of the biggest advantages is that it could help to change the way investors consider potential outcomes from PE portfolios. “If you present the outcomes as a range, it really helps people think more probabilistically, as opposed to assuming what you are presenting is a prediction,” he says.
However, in practice, even the new model has limitations. “Data can be a big limiting factor,” explains Sherwood. “An investor like Yale or GroveStreet could use this because it has been a consistent investor with decades of good-quality cash flow data to draw on. However, others might have to rely on externally sourced data and there are currently very few clean, high-quality market or industry-wide cash flow datasets available commercially that stack up to those from groups like Yale or GroveStreet. So, you may still have the problem of garbage in, garbage out, and I would be concerned if investors used this naively.”
The other issue is to do with history having a habit of not quite repeating itself, as Billias also outlines. “Like all models, you are relying on historical data and things change in a way that history can’t capture,” says Sherwood. “The dotcom bust was very different from the recent rout in technology stocks and VC portfolios because this time around, these were real companies. It’s also the case that PE is now in its first true inflationary environment since becoming a ubiquitous asset class – it hasn’t had to cope with such rapidly increasing interest rates before.”
All of this means that the new model still requires a level of sophistication among LPs. However, Sherwood says it could be valuable to LPs for forecasting cash flows and in making asset allocation decisions. “It can help investors understand the liquidity profile of a given portfolio,” he says. “It could be powerful in helping investors understand, for example, how much they should commit today if they have a target allocation of a certain level within a certain time frame. And, as investors are dealing with slower distributions today, it could help them understand what the impact might be of using the secondaries market to gain liquidity.” He adds that it is also a useful tool for helping PE teams to communicate to others the possible year-to-year variations in the asset class. Overall, he says, the new model is “a great improvement; it just needs to be used responsibly.”
Patrick Sherwood